Physical and functional interaction between SET1/COMPASS complex component CFP-1 and a Sin3S HDAC complex in C. elegans

Abstract

The CFP1 CXXC zinc finger protein targets the SET1/COMPASS complex to non-methylated CpG rich promoters to implement tri-methylation of histone H3 Lys4 (H3K4me3). Although H3K4me3 is widely associated with gene expression, the effects of CFP1 loss vary, suggesting additional chromatin factors contribute to context dependent effects. Using a proteomics approach, we identified CFP1 associated proteins and an unexpected direct link between Caenorhabditis elegans CFP-1 and an Rpd3/Sin3 small (SIN3S) histone deacetylase complex. Supporting a functional connection, we find that mutants of COMPASS and SIN3 complex components genetically interact and have similar phenotypic defects including misregulation of common genes. CFP-1 directly binds SIN-3 through a region including the conserved PAH1 domain and recruits SIN-3 and the HDA-1/HDAC subunit to H3K4me3 enriched promoters. Our results reveal a novel role for CFP-1 in mediating interaction between SET1/COMPASS and a Sin3S HDAC complex at promoters.

Figure 1.

The conserved SIN3 complex copurifies with CFP-1 and WDR-5.1. (A) Co-IP of CFP-1, WDR-5.1, ASH-2 and DPY-30. Western blot analysis of CFP-1::GFP (left panel) and HA::WDR-5.1 (right panel) purified complexes from embryos. For CFP-1::GFP, immunoprecipitations were performed on 4 mg total protein extract, and loadings were 1/200 of total protein extract for input and 1/2 of total elution volume. For HA::WDR-5.1, samples used for mass spectrometry analysis were loaded as follows: 1/8000 of total protein extract for input and 1/250 of elution for α-ASH-2; 1/60 000 of total protein extract for input and 1/200 of elution for α-DPY-30. Control samples were prepared from a wild-type strain without either transgene. (B) List of selected proteins identified by mass spectrometry of CFP-1::GFP or HA::WDR-5.1 immunoprecipitations, and their mammalian homologue. Subunits specific to the SET-2/SET1, SET-16/MLL, SIN3S and NSL complexes are highlighted in blue, green, orange and yellow, respectively. SET1/MLL core complex subunits are highlighted in grey. HCF-1 copurifies with both SET-2/SET1 and SET-16/MLL complexes. SC; Spectral Counts. (C) Cartoon representation of SET-2/SET1, SET-16/MLL and SIN3S complexes; subunits are highlighted as in (B). (D) Co-IP of CFP-1, HDA-1, SIN-3 and MRG-1. CFP-1::GFP immunoprecipitations were carried out on 5 mg total protein and analyzed by western blot using anti-HDA-1, SIN-3 and MRG-1 antibodies. Loadings were 1/5000 of total protein extract for input and 1/5 of total elution volume for anti-GFP and anti-HDA-1, 1/1000 of total protein extract for input and 1/5 of total elution volume for anti-SIN-3 and anti-MRG-1.

Figure 2.

Interaction mapping between subunits of the SET-2/SET1 and SIN3S complexes by Y2H. (A) An interaction matrix was obtained by cross-mating of yeast haploid strains expressing different subunits as AD or BD fusion proteins. Positive control (+) is barnase/barstar interaction (133). Matings that lead to visually detectable staining in two independent experiments are reported in the tabular form on the right. Failure to detect the DB DPY-30/AD ASH-2 interaction is most likely due to DPY-30 homodimerization in the context of the DB domain fusion interfering with ASH-2 binding. (B) Interaction matrix of full length and truncated CFP-1 tested against full length SIN-3 and ATHP-1. CFP-1 truncations were constructed as BD and AD fusions, as indicated. (C, D) Interaction matrix of full length CFP-1 tested against N- and C-terminal fragments of SIN-3 (C) and ATHP-1 (D).

Figure 3.

Loss of function of SET1/COMPASS and SIN3S complex components results in similar phenotypes and affects steady state gene expression. (A) Total number of progeny of single and double mutant animals of the given genotype. Multiple comparison was done using Wilcoxon (with Bonferroni correction) post hoc method following a significant Kruskal Wallis test; asterisks indicate a significant difference: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. (B) Embryonic lethality of single and double mutant strains; n = total number of embryos scored. (C) Fertility assays of set-2, cfp-1, sin-3 and athp-1 mutants grown at 25°C. Scoring was based on three to four biological replicates, with 6 independent lines each. Wildtype animals can be maintained for more that 40 generations without loss of fertility. (D) Confocal images of DAPI-stained intestinal nuclei from young adults. Examples of nuclear division abnormalities giving rise to thin chromatin bridges (arrows, top panel), or thick chromatin dense regions connecting two nuclei (arrowheads, bottom panel). Box plots show the total number of segregation defects (thin and thick chromatin bridges) per animal in single and double mutants of the given genotype (n = 150 worms for each strain). Multiple comparison was done using Wilcoxon (with Bonferroni correction) post hoc method following a significant Kruskal Wallis test as in (A). (E) Venn diagram showing the overlap between cfp-1, set-2 and sin-3 downregulated and upregulated genes. P-value for overlap between commonly misregulated genes in pairwise comparisons was calculated using Fisher's exact test in R: **P-value < 10⁻¹⁰ and ***P-value < 10⁻¹⁰⁰.

Figure 4.

H3K4me3 at strong and weak COMPASS targets is dependent on CFP-1 and SET-2. (A) IGV browser view showing z-score BEADS normalized ChIP-seq signals of CFP-1::GFP, and H3K4me3 in wildtype, cfp-1, and set-2 mutant embryos normalized to C. briggsae spike-in (see Methods). Top track shows locations of strong (S) and weak (W) COMPASS targets. (B) Heatmap of CFP-1::GFP, and H3K4me3 in wildtype, cfp-1, and set-2 mutant embryos, using same tracks as in (A). (C) Quantification of H3K4me3 signal (C. briggsae normalized) in strong and weak COMPASS targets.

Figure 5.

SIN-3 and HDA-1 require CFP-1 for recruitment to strong COMPASS targets. (A) IGV browser view showing z-scored BEADS normalized ChIP-seq signals from mixed embryos of indicated strains. Top track shows locations of strong (S) and weak (W) COMPASS targets. (B) Heatmap of z-scored BEADS normalized ChIP-seq signals from mixed embryos over strong and weak COMPASS targets in indicated strains. (C) Venn diagram showing overlap of CFP-1, SIN-3, HDA-1 and MRG-1 ChIP-seq peaks. (D) Quantification of normalized z-scored SIN-3 and HDA-1 signal in wildtype and cfp-1 mutants at strong and weak COMPASS targets.